Damage detection in the presence of outliers based on robust PCA

Abstract

Identification of outliers in samples from univariate and multivariate populations has received considerable attention over the past few decades. Presence of outliers has undeniable effects on the results of statistical methods such as Principal Component Analysis (PCA). Outliers, anomalous observations, can affect the variance and covariance as vital parts of PCA method. In statistical sense outliers are samples from a different population than the data majority. An effective way to deal with this problem is to apply a robust, i.e. not sensitive to outliers, variant of PCA. In this work a robust PCA method is used instead of classical PCA in order to construct a model using data with outliers to detect and distinguish damages in structures. The comparisons of the results shows that, the use some indexes based on the robust model,can distinguish the damages much better than using classical one, and even in many cases allows the detection where classic PCA is not able to discern between damaged and non-damaged structure. This work involves experiments with an aircraft turbine blade using piezoelectric transducers as sensors and actuators and simulated damages.